The goal of this proposal is to help achieve a unified model of implicit and explicit memory phenomena and memory distortions that span higher level conceptual and lower level perceptual effects. SAC, which stands for Source Of Activation Confusion, will be tested and challenged in several ways. First, novel predictions (i.e., not necessarily expected by other accounts) are made that could disconfirm SAC, and second, each data set will be modeled at a detailed level using SAC. Although SAC has already achieved some success in explaining and fitting data in some domains, there remain many issues to explore that were not addressed during the previous funding period. The goal is to test predictions at higher and lower levels of description. These include testing hypotheses concerning how perceptual features are integrated with conceptual features representations, and how representations change with experience, i.e., how and when higher level structures are built, and whether SAC's memory representation and processes work in an analogous fashion for elaborations (conceptual processing) and for perceptual features (data driven processing). In sum, the goals are to further test and attempt to disconfirm SAC by (a) challenging the model to account for the wide range of cross-over dissociations in the implicit/explicit memory domain, as well as memory confabulations; (b) test the model in domains where purported results already challenge SAC's account; (c) develop a more detailed representation and a more completely specified process model of encoding, matching and retrieval; and (d) shed light on aspects of cognitive functioning that will constrain theory building (within SAC and other cognitive theories). This theoretical advance will be important in helping to understand human memory, both its strength and fallibilities.